Role Summary
Owns technical testing of agent outputs — accuracy, turnaround-time impact, and regression testing as agents are iterated — feeding the accuracy and completion-time KPIs tracked at the programme level.
Key Responsibilities
• Build test suites validating agent outputs against golden-record datasets across covered processes.
• Track and report accuracy/turnaround-time metrics to support baselining exercises agreed with the business.
• Run regression testing whenever agent/prompt configurations change, to prevent silent accuracy drift in regulated processes.
• Feed findings into the human QA function for final sign-off — this role supports human QA authority on regulated activity with technical evidence, not replaces it.
Required Experience
• 5+ years QA/test engineering experience, ideally including ML/LLM output validation (not just deterministic software testing).
• Experience defining test datasets and accuracy benchmarks for document/text-processing systems.
• Domain process familiarity a plus for building realistic test scenarios.